Constrained motion field estimation for hardware efficiency

An apparatus of decoding a current block of a current frame, the apparatus comprising:

• a processor configured to execute instructions stored in a non-transitory memory to:
• select reference blocks of a first reference frame, wherein the reference blocks constitute a subset of blocks of the first reference frame;
• determine temporal motion vector candidates for the current block using respective motion vectors of the reference blocks by instructions to:
• project a reference block of the reference blocks onto the current frame to identify the current block,
• wherein the reference block is inter-predicted from a second reference frame using a first motion vector, and
• wherein the projecting uses
• the first motion vector,
• a first difference between display orders of the current frame and the first reference frame, and
• a second difference between display orders of first reference frame and the second reference frame;
• add the temporal motion vector candidate to a motion vector candidate list;
• select a motion vector from the motion vector candidate list;
• generate a prediction block using the selected motion vector;
• and
• decode, from an encoded bitstream, the current block using the prediction block.

An apparatus for decoding an encoded block of an encoded frame processed in processing units, wherein a processing unit of the encoded frame includes the encoded block, the apparatus comprising:

• a processor configured to execute instructions stored in a non-transitory memory to:
• select motion vectors corresponding to blocks of an extended collocated processing unit in a first reference frame;
• identify a block of the blocks, such that the encoded block is a projection, using a motion vector of the block that refers to a third reference frame, onto the encoded block wherein the projection comprises a scaling of the motion vector of the block using respective display orders of the encoded frame, the first reference frame, and the third reference frame;
• on condition that a block is identified,
• determine a temporal motion vector candidate for the encoded block in a second reference frame using a motion vector of the identified block and respective display orders of the encoded frame, the first reference frame, the second reference frame, and the third reference frame, and
• add the temporal motion vector candidate to a motion vector candidate list;
• select a motion vector from the motion vector candidate list;
• generate a prediction block using the selected motion vector;
• and
• decode the encoded block using the prediction block.